JP2017129564A - Abnormality determination device, abnormality determination system, and abnormality determination method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To make it possible to efficiently use a signal in an effective band to improve the accuracy in detecting abnormality of the apparatus.SOLUTION: There is provided an abnormality determination device 2 comprising: a sound collection data analysis unit 20a that analyzes data of the sound of a load of an image forming apparatus 1 collected by a plurality of types of sensors S respectively collecting sound data with different bands; a sensor selection unit 20d that selects the sensor S that has collected sound data expressing the strongest feature on the basis of a result of analysis from the sound collection data analysis unit 20a; and an abnormality determination unit 20c that determines whether the apparatus is normal or abnormal on the basis of the sound data output from the sensor S selected by the sensor selection unit 20d.SELECTED DRAWING: Figure 4

Description

  The present invention relates to an abnormality determination device, an abnormality determination system, and an abnormality determination method.

An apparatus having a load driven by a motor or the like (herein, simply referred to as an apparatus), for example, an image forming apparatus, detects an abnormal operation using an audible band signal (sound data; about 20 to 20 kHz). The technology is conventionally known.
However, the conventional technique has a problem that the band of the signal to be used is narrow, and there is a case where the signal in the band that is originally effective (which can improve the abnormality detection accuracy) cannot be used.

With regard to this point, for example, Patent Document 1 (Japanese Patent Application Laid-Open No. 2010-54558) uses a sound collected by a condenser microphone for the purpose of improving the safety of the image forming apparatus. A method for detecting an abnormality is disclosed.
This anomaly detection method uses sound collected by a microphone (audible signal), but uses a signal band that is narrow and useful (improves the accuracy of anomaly detection). The problem of not doing is not solved.

  The present invention has been made in view of the above-described conventional problems, and an object of the present invention is to make it possible to efficiently use a signal in an effective band in order to improve an abnormality detection accuracy in a device. That is.

  The present invention includes a receiving unit that receives physical data of a device acquired by a plurality of types of sensors having different bandwidths of acquired physical data, a physical data analyzing unit that analyzes the acquired physical data, and each operation mode of the device. Storage means for storing information indicating a sensor that is determined to be characterized by the physical data as a result of analysis by the physical data analysis unit when the device operates in the operation mode among the plurality of types of sensors And a sensor selection unit that selects physical data output from the plurality of types of sensors based on information indicating a sensor whose characteristics appear in the physical data stored in the storage unit, and an output from the sensor selected by the sensor selection unit And an abnormality determination unit that determines an abnormality in the operation of the device based on the operation mode based on the physical data that has been generated. A.

  According to the present invention, it is possible to efficiently use a signal in an effective band in order to improve abnormality detection accuracy in a device.

It is a figure which illustrates roughly the basic composition for the abnormality detection by the abnormality determination apparatus which concerns on embodiment of this invention. 1 is a block diagram schematically showing a configuration of an image forming apparatus. It is a block diagram which shows roughly the structure of the computer which performs abnormality determination. 1 is a diagram illustrating an abnormality determination system configured by combining an abnormality determination apparatus (computer) with an image forming apparatus. The result of analyzing the signal (sound data) of each sensor when the image forming apparatus is operated in the operation mode 1 is analyzed by the sound data analysis unit. It is a graph showing a frequency analysis result. It is a flowchart which shows the process sequence of the sensor selection for every operation mode 1, 2, .... It is a flowchart which shows the process sequence of the abnormality determination which an abnormality determination part performs using the result of a sensor. It is the same graph as FIG. 4 which shows the result of having analyzed the sensor output frequency by the selected sensor. It is a flowchart explaining the process sequence for narrowing down a zone | band.

Before describing embodiments of the present invention, the features of the present invention will be described first.
That is, according to the present invention, when detecting an abnormality in a device, for example, an image forming apparatus, different types of signals (air vibration, solid vibration, etc.) and sound data in various bands (herein, collectively referred to as “sound data”). Is characterized by automatically selecting an efficient signal type and band in order to detect anomalies.

Next, a basic configuration for abnormality detection by the information processing apparatus according to the embodiment of the present invention will be schematically described.
FIG. 1 is a diagram schematically illustrating a basic configuration for abnormality detection by an abnormality determination device according to an embodiment of the present invention.
A basic configuration for detecting an abnormality by the abnormality determination device includes an abnormality determination unit A and a signal acquisition unit B as shown in the figure.
The signal acquisition unit B is provided with a plurality of types of sensors B1, and the signal from the sensor B1 is amplified by the amplifier B2, and the abnormality determination unit A determines whether the signal is normal or abnormal. .

Here, sensors S1... SN having different bands of sound data to be collected can be used according to the situation. However, in the embodiment described below, a capacitor microphone (20 [Hz] to 12 k [Hz] , Hereinafter referred to as sensor S1), acceleration sensor (12k [Hz] to 30k [Hz], hereinafter referred to as sensor S2), and AE sensor (acoustic emission; 30k [Hz] to 1M [Hz], hereinafter referred to as sensor S3). A case will be described as an example.
In addition, each sensor S1-S3 shall acquire the one-dimensional waveform of a different band. The sensors S1 to S3 are built in the device that is the measurement target, and acquire sound data when the measurement target device operates.

Next, the case where the basic configuration for abnormality detection by the abnormality determination apparatus described above is applied to an image forming apparatus will be described as an example.
Here, an outline of the image forming apparatus will be described first.
FIG. 2 is a block diagram schematically showing the configuration of the image forming apparatus 1.
As shown in the figure, the image forming apparatus 1 schematically includes a control unit 10, an engine unit 12, an operation display unit 14, and a signal acquisition unit 16.

  The control unit 10 includes a CPU (Central Processing Unit) 101 and a ROM (Read Only Memory) 102 that stores a program for operating the CPU 101, a RAM (Random Access Memory) 103 that provides a program work area, control, timing, and the like. Based on the input of each sensor in the engine unit 12 of the image forming apparatus 1 and the non-volatile RAM 104 that can store the adjustment values and registered copy mode settings and retain the data even when the power is turned off. 12, an IO (input / output) control unit 105 that controls loads (1 to n) and the like, and a communication I / F (interface) unit 106 that communicates with a computer 2 (FIGS. 3 and 4) to be described later. Etc.

The engine unit 12 performs image formation and output of the formed image.
The operation display unit 14 inputs an operator instruction to the image forming apparatus 1 and displays the state of the image forming apparatus 1.
The signal acquisition unit 16 is disposed in the vicinity of the loads (1 to n) 121 to 124 in the engine unit 12, acquires the operation sound, and outputs the operation sound data to the connected computer 2.
Note that each of these parts is a well-known technique and is not a part having special characteristics in the present invention, and therefore detailed description thereof is omitted.

FIG. 3 is a block diagram schematically showing the configuration of the computer 2 that performs abnormality determination.
As shown in the figure, the computer 2 is roughly composed of a control unit 20 and a signal processing unit 21.

  The control unit 20 includes a CPU 201 and a ROM 202 that stores a program for operating the CPU 201, a RAM 203 that provides a work area for the program, an HDD 204 that can retain data even when the power is turned off, and the image forming apparatus 1. A communication I / F unit 205 that communicates with the communication I / F unit 106 (FIG. 2), an operation display unit 22 that displays an input instruction to the computer 2, an abnormality determination result, and the like.

  The signal processing unit 21 adjusts the level of the operation sound (analog signal) of the image forming apparatus 1 acquired by the signal processing unit 16, the signal acquisition unit 16 (sensors S <b> 1 to S <b> 3) disposed inside the image forming apparatus 1, The external I / F unit 24 that performs conversion into a digital signal is configured.

FIG. 4 is a diagram illustrating an abnormality determination system configured by combining the image forming apparatus 1 with an abnormality determination device (here, a computer) 2.
In the drawing, the image forming apparatus 1 shows only the main part of the hardware configuration shown in FIG. 2, and the computer 2 similarly shows only the main part configuration shown in FIG.

The load control unit 10a shown in the control unit 10, the sound collection data analysis unit 20a, the table creation unit 20b, the abnormality determination unit 20c, and the sensor selection unit 20d shown in the control unit 20 are all included in the control units 10 and 20. This is function realizing means that is realized by operating a CPU by a program.
The load control unit 10a performs, for example, drive control of each load (1 to n) of the engine unit 12 in accordance with the selection of the operation mode on the computer 2 side. The sound collection data analysis unit 20a performs sound collection instructions, frequency analysis of sound collection data, and the like. The table creation unit 20b creates a table showing the relationship between the operation mode and the used sensor, or the used band based on the frequency analysis of the sound collection data. The abnormality determination unit 20c determines the presence or absence of abnormality based on the sound collection data. As will be described later, the sensor selection unit 20d selects one type of sensor from a plurality of types of sensors having different bands of sound data to be collected.

Here, the control unit 10 of the image forming apparatus 1 and the control unit 20 of the computer 2 are connected via respective communication control I / Fs.
The signal acquisition unit 16 of the image forming apparatus 1 is connected to the control unit 20 via the external I / F unit 24 of the computer 2, and as described above, the signal acquisition unit 16 in the engine unit 12 of the image forming apparatus 1. Arranged in the vicinity of the load (1 to n), the operation sound is acquired by the sensor S, and the operation sound data is output to the computer 2 via the external I / F unit 24 as the reception unit.
The computer 2 analyzes the data of the operation sound collected by the sound collection data analysis unit 20a, here the frequency analysis, the table creation unit 20b creates a table to be described later, and the abnormality determination unit 20c performs the operation. Based on the sound data (for example, by comparing with normal operation sound data), it is determined whether or not each load (1 to n) is abnormal.

[Embodiment 1]
Next, a method for improving abnormality detection accuracy in abnormality determination in the image forming apparatus 1 will be described.
In the present embodiment, for example, a technique described in JP-A-2006-105267 is used as an abnormality detection method. That is, the horizontal axis represents the frequency, the vertical axis represents the intensity of the sound data of that frequency, and the amount of change (difference) between the normal time and the abnormal time is viewed. Further, the amount of change is further integrated, and when the integrated value exceeds a predetermined threshold value (the sound becomes louder / smaller as a whole), it is determined that there is an abnormality. However, the present invention is not limited to this example, and any method that acquires integral data over a frequency band and makes an abnormality determination or a method that makes a comprehensive determination using individual determination results of each frequency band is applicable. .

  In the present embodiment, three types of sensors S1 to S3 having different bands of sound data to be collected are used, and one of the most effective sensors S is selected from them, and only signals that can be acquired from the sensors S are used. Abnormality judgment is performed. When abnormality detection is performed by the method as described above, it is more accurate to perform abnormality determination using only the most effective sensor than to perform abnormality determination using signals acquired from all the sensors S1 to S3. Because it becomes higher. Here, first, the reason for selecting the most effective sensor S will be described.

That is, when a signal during normal operation and a signal during abnormal operation are compared in the frequency domain, it is considered that there is no difference in the characteristics in most bands. Therefore, if information of all bands is used, a lot of information (noise) unnecessary for abnormality determination is mixed. Therefore, as a result, the S / N ratio is deteriorated, which becomes a factor of lowering the final abnormality determination accuracy.
On the other hand, if unnecessary information is not used, the S / N ratio is inevitably improved, so that the abnormality determination accuracy can be improved. Here, the reason why a single sensor is selected from a plurality of sensors without using only one sensor from the beginning is to know in advance which sensor signal can be used effectively. Because there is no.

  Therefore, as a preliminary preparation, the sound collection data analysis unit 20a of the computer 2 analyzes the signals of the sensors S1 to S3 when the image forming apparatus 1 is operated in a specific operation mode in advance. In this case, the image forming apparatus 1 can make detailed designations for specific operations such as “mode for moving only the motor 1 (hereinafter, operation mode 1)” and “mode for moving only the motor 2” in the debug mode. .

  Based on the instruction input from the operation display unit 22, the computer 2 sends an instruction to the image forming apparatus 1 to execute the operation mode 1 via the communication I / F unit 205. Then, the sound data of the motor 1 acquired by the signal acquisition unit 16 when operating in the mode 1 is stored in the HDD 204 as normal data.

5A to 4C show the results obtained by analyzing the signals (sound data) of the sensors S1 to S3 when operated in the operation mode 1 by the sound collection data analysis unit 20a, the power on the vertical axis and the frequency on the horizontal axis. Is a graph showing the relationship between the two (frequency analysis results).
As shown in the figure, it can be seen that the strongest feature (portion surrounded by a frame) appears in the sensor S2 of FIG. 5B. Note that “a strong feature appears” is defined here as meaning that “the power is larger” than the average power when compared with the overall average power (output). This is because each load such as a motor of the image forming apparatus 1 generates a sound by operation. Therefore, it can be considered that the frequency band showing a value larger than the average power indicates the frequency band in which the sound due to the load is generated. Therefore, it may be considered that “a strong feature appears” also in a portion surrounded by a broken line in FIG. 5A. Further, “the strongest feature appears” means that the power is the highest among the “strong features”. In addition, since the range of signals that can be acquired is different for each of the sensors S1 to S3, it is necessary to perform correction as necessary so that the sensors S1 to S3 can be compared on the same scale.

  From the above results, when the image forming apparatus 1 is actually operated and the operation is in the operation mode 1, the abnormality determination is performed using the signal of the sensor S2. This method is based on the premise that when an abnormality occurs, the most obvious change appears in the band where strong features appear.

Next, a processing procedure for selecting the sensor S for each of the operation modes 1, 2,.
FIG. 6 is a flowchart showing a processing procedure for selecting the sensor S in each of the operation modes 1, 2,.
That is, first, the image forming apparatus 1 is operated in a desired operation mode (step S101), and the sound collection data analysis unit 20a acquires signals during operation by the sensors S1 to S3 (step S102), and the sensors S1 to S3. The acquired signal is subjected to frequency analysis (the power (sensor output) is examined for each frequency) (step S103). At this time, the sensor selection unit 20d selects the sensor that exhibits the strongest feature (step S104).
Next, the table creation unit 20b creates a table for each operation mode (step S105), and when the operation mode table is completed (step S106, Yes), this process ends. The table is created by storing it in the HDD 204 as a storage means.

Table 1 is a table showing the operation mode table created in this way.
Here, one sensor S to be used for each operation mode is displayed.
That is, the sensor S2 is displayed for the operation mode 1, the sensor S3 is displayed for the operation mode 2, the sensor S1 is displayed for the operation mode 3.

Next, the abnormality determination processing procedure performed by the abnormality determination unit 20c using the result of the sensor S will be described.
FIG. 7 is a flowchart showing a processing procedure of abnormality determination performed by the abnormality determination unit 20c using the result of the sensor S.
That is, first, the image forming apparatus 1 is turned on (step S201). In this state, the control unit 20 of the computer 2 extracts the operation mode from the table in Table 1 as needed (Step S202). Next, the sensor S used in the operation mode extracted from the table of Table 1 is selected (Step S203), and the abnormality determination unit 20c performs abnormality determination using only the signal acquired from the selected sensor S (Step S203). S204). After the abnormality is determined, this process is terminated.

  In this embodiment, the table is created so as to select only the sensor having the strongest feature. However, a plurality of sensors that output a predetermined amount of power larger than the average power may be selected. .

[Embodiment 2]
Next, Embodiment 2 will be described.
In the second embodiment, after selecting one of the most effective sensors S from the three types of sensors S1 to S3, a process for further narrowing the band to be used is employed.
The second embodiment may be able to acquire a signal with a further improved S / N ratio compared to the first embodiment. However, on the other hand, since the bias of information to be used is too large, there is a possibility that the abnormality determination accuracy is lower than in the first embodiment. Therefore, it is necessary to select a better abnormality determination accuracy according to the usage situation.

FIG. 8 is a graph similar to FIG. 5 showing the result of frequency analysis of the sensor output by the selected sensor S.
A method for narrowing the band to be used will be described with reference to FIG.
In the figure, it can be seen that strong features appear in the three bands (Bn1, Bn2, Bn3) surrounded by a frame.
Here, the definition of the strong feature is the same as in the first embodiment. Therefore, abnormality determination is performed using only signals in these three bands (Bn1, Bn2, Bn3).

In the present embodiment, three bands (parameters) are used. However, the present invention is not limited to this, and the number of bands (parameters) can be arbitrarily set.
As described in the first embodiment, when all the plurality of sensors that output a predetermined amount of power larger than the average power are selected, the above-described procedure is performed on all the selected sensors.

Next, a processing procedure for narrowing the bandwidth in the present embodiment will be described.
9A and 9B are flowcharts for explaining the processing procedure for narrowing the bandwidth.
First, in FIG. 9A, the image forming apparatus 1 is operated in a desired operation mode (step S301), the sensor S shown in the table is selected, and a signal (sound collection data) during the operation by the selected sensor S is selected. ) Is acquired (step S302). Next, frequency analysis is performed on the acquired signal (step S303), and the sensor S in which the strongest feature appears is selected (step S304). Next, the band in which strong features appear (for example, Bn1, Bn2, Bn3 in FIG. 8) is narrowed down (step S305), and a table for each operation mode is created (step S306). If the table for each operation mode is completed (step S307, Yes), this process ends.

  After the table for each operation mode is completed, when the image forming apparatus 1 is turned on in FIG. 9B (step S401), the operation mode is extracted from the table table as needed (step S402). Next, a sensor and a band to be used for each operation mode extracted from the table are selected (step S403), and abnormality determination is performed using only the signal acquired from the selected sensor and band (step S404). finish.

Table 2 shows an example of a table for each operation mode.
Here, the sensor S used for each operation mode and the band used are displayed.
That is, for the operation mode 1, the sensor S2 and the use band are 16 k [Hz] to 18 k [Hz], and for the operation mode 2, the sensor S3 and the use band are 300 k [Hz] to 400 k [Hz]. , 500 k [Hz] to 600 k [Hz], for operation mode 3, the sensor S1, the use band is 1 k [Hz] to 2k [Hz], 4k [Hz] to 5k [Hz], 10k [Hz] ~ 11k [Hz].
In this way, by narrowing the use band of sound data when performing abnormality determination, it is possible to perform abnormality determination with higher accuracy than when using the entire use band of sound data.

  As described above, the embodiment of the present invention has been described. However, according to the present embodiment, abnormal signals can be obtained by using different types of signals (air vibration, solid vibration, etc.) and multiple types of sensors that can acquire various bands. Is detected. For this reason, since an efficient signal type and band are automatically selected, it is possible to efficiently use a signal in an effective band in order to improve anomaly detection accuracy.

In the embodiment of the present invention, the image forming apparatus is taken as an example of the abnormality determination target device, but the present invention is not limited to this. Any device that performs a predetermined operation in accordance with a predetermined mode and generates sound in accordance with the operation can be applied.
Further, the sound data is not limited to this, and the frequency of vibration generated according to the operation or the acceleration accompanying the vibration may be detected as the frequency, and data indicating any of these physical quantities may be acquired and used. .

  DESCRIPTION OF SYMBOLS 1 ... Image forming apparatus, 10 ... Control part, 10a ... Load control part, 101 ... CPU, 102 ... ROM, 103 ... RAM, 104 ... Nonvolatile RAM, 105 * ..IO control unit 106 ... Communication I / F unit 12 ... Engine unit 14 ... Operation display unit 16 ... Signal acquisition unit 2 ... Computer 20 ... Control Unit, 20a ... sound collection data analysis unit, 20b ... table creation unit, 20c ... abnormality determination unit, 20d ... sensor selection unit, 22 ... operation display unit, 24 ... external I / F part.

JP 2010-54558 A

Claims (11)

A receiving unit that receives physical data of devices acquired by a plurality of types of sensors having different bandwidths of physical data to be acquired; and
A physical data analysis unit for analyzing the acquired physical data;
Of the plurality of types of sensors for each operation mode of the device, information indicating a sensor that is determined to be characterized by the physical data as a result of analysis by the physical data analysis unit when the device operates in the operation mode Storage means for storing
A sensor selection unit that selects physical data output from the plurality of types of sensors based on information indicating a sensor characterized by physical data stored in the storage unit;
An abnormality determination device comprising: an abnormality determination unit that determines an abnormality in operation of the device according to the operation mode based on physical data output from a sensor selected by the sensor selection unit. In the abnormality determination device according to claim 1,
The physical data analysis unit performs frequency analysis of the physical data,
The abnormality determination apparatus, wherein the storage unit stores information indicating a sensor that is characterized by the physical data. In the abnormality determination device according to claim 1 or 2,
The physical data analysis unit performs frequency analysis of the physical data,
The storage means stores information indicating a sensor that acquires physical data of a band having a value larger than an average among the physical data as information indicating a sensor that is characterized by the physical data. apparatus. In the abnormality determination device according to any one of claims 1 to 3,
The abnormality determination apparatus, wherein the storage unit stores information indicating a sensor characterized by the physical data for each operation mode of the device. In the abnormality determination device according to any one of claims 1 to 4,
The physical data analysis unit further performs frequency analysis of the physical data,
The abnormality determination device, wherein the storage unit stores a band having a value larger than an average among the physical data.   An abnormality determination system including the apparatus connected to the abnormality determination device according to claim 1 to which a plurality of sensors for acquiring the physical data are attached, and the abnormality determination device. A receiving step of receiving physical data of the device acquired by a plurality of types of sensors having different bandwidths of physical data to be acquired;
A physical data analysis step for analyzing the acquired physical data;
Of the plurality of types of sensors for each device operation mode, information indicating a sensor that is determined to have a feature in the physical data as a result of analysis by the physical data analysis step when the device operates in the operation mode. Storing in the storage means;
A sensor selection step of selecting physical data output from the plurality of types of sensors based on information indicating a sensor characterized by the physical data stored in the storage step;
An abnormality determination method, comprising: an abnormality determination step for determining an abnormality in operation of the device according to the operation mode based on physical data output from the sensor selected in the sensor selection step. In the abnormality determination method according to claim 7,
The physical data analysis step performs frequency analysis of the physical data,
In the storage step, information indicating a sensor whose characteristics appear in the physical data is stored. In the abnormality determination method according to claim 7 or 8,
The physical data analysis step performs frequency analysis of the physical data,
In the storage step, information indicating a sensor that acquires physical data in a band having a value larger than an average of the physical data is stored in the storage unit as information indicating a sensor characterized by the physical data. An abnormality determination method. In the abnormality determination method according to any one of claims 7 to 9,
The storage step stores in the storage means information indicating a sensor characterized by the physical data for each operation mode of the device. In the abnormality determination method according to any one of claims 7 to 10,
The physical data analysis step further performs frequency analysis of the physical data,
The storage step stores in the storage means a band having a value larger than an average among the physical data.

Images